Bart Wessel, the Hagelin Cryptographers C-52 and CX-52
Total Page:16
File Type:pdf, Size:1020Kb
February 24, 2021 The Hagelin Cryptographers C-52 and CX-52 V1.01 Update March 12, 2021 The Hagelin Cryptographers C‐52 and CX‐52 Bart Wessel Keywords: Hagelin; Crypto AG; Cryptoteknik; C-52; CX-52 Abstract While the C-52 and CX-52 devices enter the open market and are exhibited in museums, it is hard to the untrained eye tell the differences. There is a lot of confusion, and half-truths muddle the waters even further. It turns out that there is only a limited number of standard configurations of C-52 and CX-52 machines. These ‘models’ are described in detail, to enable the reader to identify machines he/she comes across. All the configurable components are described to aid in the understanding of their differences and of the way the devices work. The C-52 is usually cryptographically limited compared to the CX-52 – although an equally limited CX-52 is entirely possible! It all seems to come down to marketing and willingness of Hagelin to sell a certain cryptologic complexity to a particular customer – with the NSA looking over his shoulder. The cryptologic quality of the various models is not explained here in mathematical detail. The author hopes to have presented a defining framework of configurable variables which can be built upon by a more crypto-analytical minded author. Introduction The Hagelin C-52 and CX-52 are deceivingly simple looking mechanical cryptographic machines that date from the early 1950’s and are familiar to many cryptography enthusiasts, collectors and museums. Crypto AG, owned by Boris Hagelin1, secretly created machines that differ in internal configuration without the customer being aware. The C-52 and CX-52 devices seem, by virtue of their names, to date from 1952. Yet in August 1950 key features of the C(X)-52 where already in development and shared with AFSA (Armed Forces Security Agency, the forerunner of the NSA). In September Friedman2 refers to the description of an apparently new Hagelin design with an internal AFSA note that starts “I think you’ll find this rather interesting” i. A 1954 NSA report ii on preliminary studies on agency COMINT needs for large-scale analytic equipment mentions that: Any CX-52 traffic that appears will tax our present facilities severely. It will require more complicated, faster programs; and most of the present special-purpose equipment will not be applicaBle. Widespread uncontrolled use of the CX-52 was clearly undesirable from the point of view of the NSA. In January 1955 there is talk of AFSA loaning a C-52 and a CX-52 from Hagelin, a little later some machines were ordered from the factory, indicating serious interest from AFSA in this development. 1 Boris Caesar Wilhelm Hagelin (1892 –1983). 2 William Frederick Friedman (1891 –1969) was a US Army cryptographer who ran the research division of the Army's Signal Intelligence Service (SIS) in the 1930s, and parts of its follow-on services into the 1950s (Wikipedia). He and Boris Hagelin became good personal friends. 1 - 30 Bart Wessel The C-52 and CX-52 form a family of machines, there is no such thing as ‘the’ C-52 or ‘the’ CX-52. A C-52 and a CX-52 can be internally configured to function identically, or two CX-52’s can be configured to perform wildly differently – and still both be labeled Type CX-52, Series D. The name plate does not reflect the actual configuration of the machine as part of the companies’ explicit rouse to keep customers uninformed. Today, this leads to confusion among collectors and curators as seemingly identical machines behave very differently. By comparing a multitude of machines, original factory brochures and spare part lists, this study aims to identify the internal configurations of C-52 and CX-52 machines. To distinguish between them, we use the term ‘models’. Some models are identified and described from documentation even when no actual machine of that configuration has been spotted yet. It seems likely that specific configured behavior may have been destined for specific customers (countries) or groups of customers (regions/ideologies) and was fully intended so by Hagelin. The C-52 and CX-52 share a lot of parts, most notably: the housings are identical. Many of the parts carry similar part numbers; C-46750 Side cover cpl. iii is identical and exchangeable with CX-46750 Side cover, complete iv. Where parts differ, the part numbers bear no relationship. Differences (mainly in the drum and the drive mechanism for the keywheels) are described in detail. 2 - 30 The Hagelin Cryptographers C-52 and CX-52 Terminology Some of the terms and concepts are introduced immediately for reasons of clarity and consistency. Also three identifiers are introduced: types, series and models. It is important to understand the differences. Terms The terms in Bold typeface have been taken from the manual Instructions for Operator v, the Spare parts catalogue vi, THE STORY OF THE HAGELIN CRYPTOS vii, or Description of CRYPTOGRAPHERS “HAGELIN” TYPE CX-52 viii. In brackets are terms that may be familiar from descriptions of other makes of machines. 1. Typewheel unit, alphabet unit, print wheels, (print head). Contains primary typewheel and secondary typewheel. Rotation is referred to as displacement in displacement steps. 2. Journal plate. Forms a complete assembly together with the guide arm and the keywheel 3. Guide arm (sensing arm). Each keywheel has its own. 4. Keywheel, key wheel, pin disk¸ (cipherwheel, pin-wheel). Rotation is referred to as advancement to distinguish between movement of the alphabet unit and the keywheels. One speaks of the arrangement of pins on a pin disk. 5. Bar drum, drum, (cage). Holds the 32 slide bars. 6. Bar, slide bar, (sideways moveable bar). The word displacement is also used in connection with the slide bars. We call the default right-hand position the undisplaced and the left-hand position the displaced position. Depending on the cams on the bar we distinguish Advance bars and Non- advance bars. 7. Cam, tooth. Cams are used to displace the typewheel unit and advance the keywheels. 8. Lug, bar lug. Lugs are placed by the user as part of the key. 9. Locking bar, top cradle. Locks the guide arms in the CX-52 and limits their movement in the C-52. 3 - 30 Bart Wessel Ciphering is used to avoid repetition of the expression ‘enciphering or deciphering’. Slide bar specification As an example, this fictional slide bar 77 is specified as follows: (left) (right) 1st 2nd 3rd 4th 5th 6th print wheel wheel wheel wheel wheel wheel Bar ID lugs K 0 A B 0 C 0 77 ñ Cams, one for printwheel and six for keywheels: K: displace print wheel when bar displaced. A: advance this keywheel when bar displaced. B: advance this keywheel when bar not displaced. C: advance this keywheel. 0: do not advance this keywheel. ñ Bar-id: specific number on the slide bar to identify it. This fictional bar (no slide bar has been identified that incorporates all the different cams) has number 77. ñ Lugs: some bars cannot be lugged. If it cannot be lugged, this is indicated by ‘No’. Slide bars and their cams are explained in extensive detail below. C(X)-52Types, series and models There are a number of publicly available documents on the C(X)-52. From these, the meaning of the type, series and model designators become clearer. These documents also give an idea of how the NSA viewed the machine at the time. This research however does not claim to be complete. Types The ‘type’ designator only distinguishes the machine from earlier ones like the C-446 and M-209. All C-52’s are labeled Type C-52 and all CX-52’s are labeled Type CX-52 on the serial number plate. Series ‘Series’ refers to the designator (none, A, B, C and D are known so far) on the serial number plate like for example Type CX-52, Series D. In the spare parts catalogue vi, the series A, B, C and D are mentioned (see for example pages 5, 15, 16). Spare part catalogues suggest a strong link between the series designator and the serial number. At least minor manufacturing differences between the series exist. The series designator appears to indicate a state of mechanical evolution during the production life cycle of the C(X)-52. 4 - 30 The Hagelin Cryptographers C-52 and CX-52 For the C-52, recognizable mechanical changes in development are: Slide bar typewheel/Index Series Serial number retaining ring Lock type wheel Series B onwards Shift lock Series A & B Cross key < 25.000 Cross key Series C onward Flat key Series A & B & C Conical gear < 25.000 Conical gear Series D onwards Straight gear For the CX-52, recognizable mechanical changes in development are: Slide bar typewheel/index Series Serial number retaining ring Lock type wheel Series A Push-pin < 21.999 Push-pin Series B onwards Shift lock Series A & B Cross key 21.101 … 21.999 Cross key Series C & D onwards Flat key Series A & B & C Conical gear 21.101 … 21.999 Conical gear Series D onwards Straight gear Notes: ñ The slide bar retaining ring can be rotated after unlocking it. In early versions, this is done by pushing down a small locking pin. With later versions, the pin is to be shifted. ñ The cross key lock is familiar from the C-446 and BC-543 machines. This type of lock was later replaced by one with a common flat key.